Table 9 Performance of the classification machine learning models in JARVIS-ML with JARVIS-DFT data using OptB88vdW (OPT) and TBmBJ (MBJ) with Receiver Operating Characteristic (ROC) Area Under Curve (AUC) metric.
Property | Number of datapoints | ROC AUC |
|---|---|---|
Metal/non-metal (OPT) | 24549 | 0.95 |
Magnetic/Non-magnetic (OPT) | 24549 | 0.96 |
High/low solar-cell efficiency (TBmBJ) | 5097 | 0.90 |
High/low piezoelectric coeff | 3411 | 0.86 |
High/low Dielectric | 3411 | 0.93 |
High/low n-Seebeck coeff | 21899 | 0.95 |
High/low n-power factor | 21899 | 0.80 |
High/low p-Seebeck coeff | 21899 | 0.96 |
High/low p-power factor | 21899 | 0.82 |
